Dynamic response optimization of a thermoplastic composite sandwich beam under random vibration

Abstract

The dynamic response of a thermoplastic composite sandwich structure is optimized under random vibration. First, the experimental modal analysis data of a set of test samples are processed by a sequential set of statistical analysis such as descriptive statistics, factor analysis, and paired sample t-test. Then, the sample with the highest ability to represent the group is taken as the reference data. Three different computational models, which are defined according to whether the solid to be meshed is considered an area or a volume, are constructed. Modal analysis results of the computational models are compared to the reference experimental data to evaluate the performance of the models. To predict the dynamic response of the sandwich beam, it is excited through a random signal in the transverse direction. The nodal acceleration responses are computed in 17 evenly spaced points located on the upper finishing layer of the sandwich beam. Finally, a geometry optimization study is conducted to predict the optimum thicknesses of the 7 layers bonded together to form the sandwich beam. The optimum layer thicknesses that minimize the nodal accelerations at 17 evenly spaced points on the sandwich beam are computed. The current study shows that the shell model has the closest values to the experimental data compared to other models. As far as the dynamic response of a TPC sandwich structure is concerned, it is concluded that the shell model better represents the structure during the modeling phase and leads to concurrently reduced weight and nodal acceleration, when optimized.