APPLICATION OF DYNAMIC KINETIC ENERGY AND ARTIFICIAL NEURAL NETWORK TO CRACK IDENTIFICATION IN BEAMS UNDER MOVING MASS

In this paper, a novel method has been proposed to identify cracks in beam structures under excitation of moving mass. For this purpose, the dynamic kinetic energy of cracked beam under moving mass were used as criteria to detect crack in structures. Dynamic kinetic energy is depending on the mass and velocities of beam structures under moving mass excitation. The dynamic kinetic energy of cracked beam has been used as input of artificial neural networks in which the crack states as output. This data is acquired by the analysis of cracked structure applying the finite element method (FEM). The artificial neural networks performance has been investigated in training, validation and testing stages. Validation is used to measure network generalization, and to halt training when generalization stops improving. Testing has no effect on training and so provides an independent measure of network performance during and after training. Runge-Kutta 4th Order method has been used to solve the equation of motions of studied beams in Matlab (2015). A validation study has been done with an example that reported in literature. To evaluate the efficiency of the proposed method, two numerical examples consisting of simply supported beam and fixed simply supported beam have been studied. To be more compatible with the real dynamic cases, another examination was performed in which the dynamic kinetic energies with 3% noise are used in crack identification. To perform this, some random noise has been added to the theoretically calculated dynamic kinetic energies. Also, the modeling errors in the analytical model have been studied. It is assumed that the actual tested beam has perturbations of stiffness of 2% at some elements. The obtained results reveal that the presented method is robust and reliable to detect cracks in beam structures under moving mass. Also, the proposed method shows good results using noisy dynamic kinetic energies data.