Multiple Damage Identification in a Beam Using Artificial Neural Network-Based Modified Mode Shape Curvature

In the present work, the existence of multiple damage locations is identified successfully by using the modified mode shape curvature technique in a cantilever beam. The noisy frequency response of the beam is extracted for varying damage depths at two various positions by using Bruel and Kjaer instrument. As experimentally obtained displacement mode shape data cannot reflect clear damage location in the structure due to the presence of noise, in the present work, the data have been trained through artificial neural network to obtain improved results to localize the damage locations. Numerically and experimentally obtained displacement modes are utilized as input for ANN, and the trained data are used to produce mode shape curvature. The trained data sets are then utilized to produce the mode shapes curvatures for all the damage cases using central difference approximation. Damage severity and locations are then identified by analyzing the absolute mode shape curvature difference for various damage scenarios.

Automated summary

The study successfully identified multiple damage locations in a cantilever beam using the modified mode shape curvature technique and artificial neural network training. Experimental and numerical displacement modes were used as input for ANN to improve the accuracy of damage location identification.