Application of Spectral Element Method and Machine Learning Technique in the Composite Concrete Beam Damage Prediction

The paper presents the simulation results of the monitoring and prediction of damage of a cement composite beam using the analysis of changes in resonant frequencies with Spectral Element Method and the Machine Learning technique. For a beam with random generated crack, the effect of cracking on the change of resonance frequencies was shown by analyzing resonance curves comparing to the health beam. For the four basic resonance frequencies, frequencies were randomly generated, constituting the frequency database for the Machine Learning—Support Vector Machine classification. Damage Index was calculated. The presented methodology is a procedure involving the simulation and prediction of beam damage.