Riveting deformation prediction of aircraft thin-walled parts based on BAS-BP neural network

Abstract The contact surface is rough, which affects the riveting process. Firstly, the mechanical analysis of riveting deformation is carried out. Force and contact stiffness of joint surfaces are constructed by fractal contact theory. The fractal contact model of the rough surface of thin-walled parts is established. Based on the measured data of the riveting surface, the finite element analysis model of the rough surface of thin-walled parts is established. The interference uniformity of thin-walled parts is influenced by variables such as riveting force, riveting force action time, the length of the rivet, thin-walled part aperture, and fractal dimension. Finally, the BAS-BP is compared with the riveting deformation prediction data of thin-walled parts optimized by other algorithms.