The more our infrastructure is aging, the risk of structural failure is higher, making the detection of damage using modal parameters a very important factor that can be applied in structural health monitoring. The most desired way to assess the health of engineering structures during operation is to use non-destructive vibration-based methods. In the current paper, a modal approach using a machine learning technique by training a feedforward backpropagation neural network for detecting transverse damages in simple supported beam-like structures is presented. A method for analytical determination of the training data is used and the obtained dataset values are employed for training an ANN that will be used to locate and evaluate the severity of transverse cracks in cantilever beams. The output from the ANN model is compared by plotting the calculated error for each case in comparison with FEM results using the SolidWorks simulation software.

Keywords: damage detection, machine learning, natural frequency, structural health monitoring