Assessing Predictive Capabilities for Nonlinear Dynamic Structural Responses

摘要

Complex assemblies subject to engineering environments often contain material, interfacial, or geometric features that produce nonlinear responses. One specific interfacial feature of interest is contact under vibrational loads. Our project expands upon previous work by developing both linear and nonlinear finite element (FE) models of a four degree-of-freedom benchmark testbed, configured for variable states of intermittent contact, using Abaqus/Explicit. The goal of our study is to explore the effects of available contact interaction models on quantitative validation metrics. A better understanding of contact algorithms in FE software would help increase confidence in FE models of complex, nonlinear dynamic systems. The validation metrics of interest include features of the acceleration time series such as basic statistics (peak amplitude, mean, standard deviation, skewness, kurtosis, normal probability plot) - as well as features of the frequency domain including natural frequencies, mode shapes, power spectral density, and frequency response function plots. First, the linear FE model is used for baseline verification and validation purposes. Once it is validated against experimental data to within an acceptable margin of error, the linear model is used as the base for the nonlinear model. The nonlinear simulation explores Abaqus contact algorithms and their effects upon the dynamic response of the structure. In particular, the contact interactions we focus on are "hard" contact and the various types of "softened" contact. Finally, the nonlinear model is validated against data collected from experiments with the physical structure and the effects of the various contact interactions are analyzed.