NUMERICAL AND EXPERIMENTAL CORRELATION OF A RECTANGULAR BEAM USING A GUYAN-SEREP MIXED METHOD

Autores

  • Claudio de Oliveira Mendonça
  • Ricardo H. Gutiérrez
  • Ulisses A. Monteiro
  • Luiz A. Vaz
  • Brenno C. Moura

Palavras-chave:

Model reduction, Guyan method, Serep method, Model correlation

Resumo

Structures and equipment under dynamic loading are susceptible to a vibration response,
which can reduce its reliability and life span. Continuous monitoring can be a complicated and expensive
task, its complexity makes it impossible to always measure vibration in the locations with higher risk of
failure, especially in regions with poor accessibility. One way of estimating vibration levels of the
structure in areas of interest is using numerical Finite Element method models, however, these models
have the disadvantage of not always representing the real structure adequately, as they do not take into
account manufacturing errors and other uncertainties. The differences between model and real structure
may be minimized through calibration using experimental data, nevertheless, another difficulty arises,
as the number of degrees of freedom in the finite element model are a lot bigger than the number of
measured degrees of freedom, therefore, calibration cannot be performed directly. The use of reduction
techniques makes the calibration, regarding the modal parameters, feasible as it allows the compatibility
between model and experimental degrees of freedom, therefore allowing prediction of vibration levels
at any point in the structure. In the present work we have used a mixed GUYAN-SEREP methodology
for the model reduction of complex structures. This process is based on two steps: first, using GUYAN
method to reduce, on the physical domain, the complete model up to a manageable number of degrees
of freedom; then, secondly, using SEREP method to end the reduction, on the modal domain, ensuring
the compatibilization of the degrees of freedom with the available experimental data. This methodology
was applied to a rectangular beam, free-free condition, the results were compared with data obtained by
an experimental modal analysis, by means of MAC, relative difference (RD) and coordinate modal
assurance criteria (COMAC), obtaining high accuracy. Finally, a correlation of numerical modes was
undertaken in relation to the experimental modes yielding improvements on the results for the criteria
used.

Downloads

Publicado

2024-08-26

Edição

Seção

Artigos