MACHINE LEARNING-BASED INVESTIGATIONS ON NONLINEAR VIBRATIONS OF CFRP COMPOSITE

Carbon fiber-reinforced polymer (CFRP) composite has been increasingly used in constructions, which is subjected to external loads that cause severe nonlinear vibrations and final structural failure. In this work, a machine learning approach is adopted to investigate nonlinear vibrations of CFRP composite. A dataset about nonlinear frequency ratios of CFRP composite under different load levels, sizes, and boundary conditions is established, which is used to train prediction model. Using developed model, attribute importance of nonlinear vibration and nonlinear frequency ratios under various conditions are investigated. The finding of this work contributes to predicting long-term vibration properties of CFRP composite.