Identification of Crack of Cantilever Beam Using Experimental Results and a Hybrid Neuro-GA Optimization Technique

Abstract

In this paper, dynamic behavior of free transverse vibration of an isotropic beam having single crack has been captured for cantilever boundary condition. Condition of the crack is granted stay open throughout the operation. The free vibration experimentation is carried out by exciting the system at its deflected configuration with the blow of a hammer of soft rubber, and the feedback is collected by applying an accelerometer mounted on the test specimen. Then, finite element model of beam with different boundary conditions with single and multiple open and breathing transverse cracks is developed in ANSYS environment. Followed by cracked beam is modeled and three-dimensional FEM analysis is implemented using ANSYS. Comparison studies of experimental result with finite element analysis are executed. Results collected by the experimentation are applied in cascade neural network, genetic algorithm, and cascade neuro-GA crack identification optimization techniques. The ‘Inverse problem’ consists of calculating the damage parameters from the frequency shifts of crack beams. The merit and demerit of these optimization techniques are focused.