Neural Based Controller for Smart Detection of Crack in Cracked Cantilever Beam

Abstract

Over the years damage detection in structures is being given prior attention. For this purpose different newer techniques are being used so far. In this paper, defect in cantilever beam in the form of a transverse crack is being investigated by developing a Neural network based controller. The input parameters to the controller are relative divergence of first three natural frequencies and the output parameters are relative crack depth and relative crack location in dimensionless forms. The rules for the controller are derived from the experimental analysis and a pool of data was created to develop the controller. The validation of the developed model was done to check its effectiveness in predicting the output responses. It was found that the developed neural network based controller can predict the location and depth of the crack in a close proximity as par with the actual results.