Dynamic modeling of bolted joints for radially segmented annular virtual material considering thickness

To improve the precision of the dynamic modeling of bolted joints, a novel method is presented to model the dynamics by using a radially segmented annular virtual material equivalent bolted joints considering the thickness in this work. Firstly, on the basis of obtaining the contact pressure distribution at bolted joints, the division criterion of the virtual material layer along the radial segments is proposed, and the finite element (FE) model of the equivalent bolted joints of the radially segmented annular virtual material considering the thickness is established. Secondly, the characteristic parameters (elastic modulus, poisson’s ratio, density, and thickness) of the radially segmented annular virtual material are used as inputs to calculate the natural frequencies of the bolted joints system, and the deep neural network (DNN) model is constructed and trained on this data sample. Finally, the cuckoo search (CS) algorithm is used to optimally identify the characteristic parameters of radially segmented annular virtual material. The bolted joints of the self-designed solid model of the bolted joints system are studied and analyzed as an example, and the maximum relative error of the first six orders natural frequencies lowers than 1.5 %.