Analysis of free vibration of stiffened plates due to variation in componet dimensions and weld profiles

Wongwitit, Sutti (2005) Analysis of free vibration of stiffened plates due to variation in componet dimensions and weld profiles. Masters thesis, Memorial University of Newfoundland.

[English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (10MB)

Abstract

The investigation, reported in this thesis, focused on the free vibration responses of two identically fabricated stiffened plates due to the inherent variation in dimensions of components of stiffened plates and also due to the presence of weld profiles along the connections between the components. The percent certainty of 99.73 of each data set to be measured was used, lying in between μ - 3σ and μ + 3σ of each data set; moreover, three sets, μ - 3σ, μ and μ + 3σ, of each dimension of the components and weld profiles were used to generate finite element analysis models. The possible effects of variation in spacings, between the two transverse girders, were analyzed by setting the spacings to be normal, close girder and away girder cases, respectively. Weld cross sections were transformed to equivalent rectangular cross sections in order to use two-dimensional quadrilateral shell elements to model in finite element analyses. The validation of the transformation was also examined. -- The finite element analysis models were categorized into two groups, the models without and with weld profiles. Each model was generated using three different spacings of the girders, and each of them was generated using three prescribed sets of dimensions. All models were broken up into 4 stages in order to consider the influence of addition of components to structural behaviour and also to observe the variations in natural frequencies. -- Two finite element analysis software packages, viz., ABAQUS and ANSYS, were used to confirm the reliability of certain results. In addition, a number of element types and element sizes were used to determine the suitability of element types and convergence of results. -- Two approximate methods were employed to obtain the natural frequencies of stiffened plates. The first one was the method of elastic equivalence, and the other one was the approximate method using concepts of static analysis. The results obtained from these two methods and finite element method were compared. -- Once the weld profile geometries of the two models were obtained, it was observed that the weld qualities of Model I were better than those of Model II because the standard deviations of widths of weld on panel and on webs of girders of Model II were 1.39 and 2.59 times greater than those of Model I; the standard deviations of all the other dimensions of both models were comparable. -- From all the analyses, it was observed that the variation in dimensions of components of stiffened plates and in weld profiles produced a variation in natural frequencies between -4.59% to +4.74% for Model I and -4.27% to +5.04% for Model II. The inclusion of weld profiles alone in the models produced a variation in natural frequencies between of -5.02 to +4.12% for Model I and -4.50 to +3.14% for Model II. Moreover, the inaccurate placement of the transverse girders produced -2.01 %to 2.04% variation in natural frequencies for Model I, and -2.03 % to 1.99 % variation in natural frequencies for Model II. The method of elastic equivalence gave differences of -68.54 % and -59.96 % for the comparable natural frequencies of this method and finite element method. In addition, an approximate method using concepts of static analysis was also used to determine the first five natural frequencies of stiffened plate. The differences in results obtained from this method and finite element method were found to be between -4.39 %to +5.99 %.

Actions (login required)

View Item