Crack prediction in beam-like structure using ANN based on frequency analysis

The dynamic experimental and numerical analysis of cracked beams has been studied with the aim quantifying influence depth crack on response steel beams. Artificial Neural Method ANN used where a simulation was improved in Matlab. A finite element model also developed by using Ansys software, obtained results were compared exact length. study takes into account different hidden layer values order to determine sensitivity predicted depth. show that beam (frequencies) is strongly related which significantly affects behavior, especially when very deep allows us identify lower computational time. Based provided results, we can detect effect size affect results.