Power spectral density moment of having defective 3D printed plastic beams under moving load based on deep learning

3D printing and 3D printing technology are increasingly popular in today’s world. However, there have not been many studies evaluating the quality of 3D printed products in real-life applications. This manuscript proposes a parameter for monitoring deterioration conditions of 3D printed plastic structures based on a multilayer perceptron network, using power spectral density (PSD) under a moving load. To create deterioration phenomena in the 3D printed plastic beam structures, simulations with cracks that change the stiffness of the structure are conducted. The features presented in this manuscript are constructed from the alteration forms of power spectral density used to detect the deterioration of a 3D printed plastic structure, accomplished by creating damage in beams and using a multilayer perceptron network in an input training dataset. Under these circumstances, the power spectral density is established by vibration signals obtained from acceleration sensors scattered along the 3D printed plastic beams. The results in this manuscript show that differences in the shapes of the PSD attributable to damage are more noticeable than those in the value of the basic beam frequency. This means that adjustments of shape in PSD will better allow the detection of damage in different 3D printed plastic beam structures. The determination of defects on 3D printed plastic beams by the power spectral density method has been used in research. However, the application of this deep learning model presents many new and positive effects.