Vibration based damage identification method for cantilever beam using artificial neural network

Abstract

This paper presents the use of Artificial Neural Networks (ANN) techniques to identify the damage in cantilever beams. We consider two cantilever beams of different materials i.e. aluminum and stainless steel. Different crack lengths are introduced on the beams from 0–10 mm with 2 mm interval. 0 mm, 2 mm, 4 mm, 6 mm, 8 mm and 10 mm cracks denotes damage level 0, 1, 2, 3, 4 and 5 respectively. The undamaged cantilever structure is treated as damage level 0. Experimental modal analysis is conducted for each case using impact hammer test. To validate the experimental values modal analysis is conducted in ANSYS software. From the modal analysis results it is observed that, for lower modes there is no change in frequencies but for higher modes the natural frequencies are decreasing with the increase in crack length. The FRFs obtained from experimental modal analysis are used as inputs to train the ANN. In the present paper, two types of networks are considered. One is Radial Basis Function (RBF) network and other is feed forward network. For each material total of 60 sets of data were collected. Part of the data is used to train the ANN and remaining data is used to test the trained ANN. From the results, the ANN is capable of identifying the damage and its severities.