Structural Health Monitoring for RC Beam Based on RBF Neural Network Using Experimental Modal Analysis

This paper presents an application using Radial basis Function with Neural Network RBF-NN to predict damage identification. Free vibration experiments at various degrees of damage analysis is done to study the complex behavior of RC beams affected by notches. Both intact and damaged states are analyzed based on experimental measurement. Different damages degrees concentrated in the middle area, a beam model with a notch is built. The frequency response function (FRF) envelope acquired by the dynamic experimental tests is established and the changes in normal frequency values are compared with the degree of damage in the models of the RC beams. The action of damaged beams under free vibration is improved by implementing analytical models with equivalent stiffness reduction and identification processes. In order to verify the damage analysis procedure based on vibration testing and the recommended analytical method, a comparison of experimental and analytical results is given as the first stage. In the second stage, the effectiveness of (RBF-NN) is provided for different damage scenarios to predict the notches depth and width of RC beam. The results showed the robustness of RBF-NN to predict the notches depth and width of RC beam with a high accuracy.