DETERMINING THE POSITION AND SEVERITY OF A TRANSVERSE CRACK IN COMPOSITE STRUCTURES USING MACHINE LEARNING

Abstract

Currently, composite materials are used more often in engineering constructions in fields such as aerospace, energy, and military as well as in the car manufacturing industry. For ensuring the safe operation of structures produced from multilayered materials, it is good practice to apply Structural Health Monitoring (SHM) methods. In the current paper, we present an SHM method based on modal analysis, for evaluating open transverse cracks that can be present in a 5-layered composite cantilever beam. To achieve the proposed objectives of the research, we use a method for calculating the Relative Frequency Shifts (RFS) caused by the damage, by expressing the severity of the crack as a function of the stored energy. The obtained RFS’s values are used as input data for training an artificial neural network (ANN), which can determine the crack's position and severity.