Numerical Simulation Of Friction Welding Process Of GH4169 Superalloy

GH4169 superalloy is widely used in high-tech fields, such as aerospace, nuclear industry and so on, which usually need excellent welding quality of superalloy parts. But the traditional fusion welding can't meet the demands. Friction welding, as one of the most advanced solid state welding, has been the main joining method of superalloy in high-tech fields. Because it is used in high-tech domain, the related technologies are blanked off in the oversea companies. So we must independently study the friction welding conditions for superalloy parts. Traditionally, new welding conditions were studied by the accumulated experience of trial and error approach, which was not suitable to the development level of our country's society and economy for it usually result in high cost and long lead-time. In order to reduce the cost and shorten the lead-time, the numerical simulation method was adopted to investigate the friction welding process, and then some experimental works were carried out to demonstrate and revise the numerical model, by which the evolutions of physical parameters during inertia friction welding were analyzed and the influence of technological parameters and geometrical parameters on the inertia friction welding process were explored. And then the ANN (artificial neural networks)prediction system of upset was also constructed for inertia friction welding of ring parts.In this paper, a 2D axisymmetric thermal-mechanical coupled FE model of inertia friction welding process was developed using elastic-plastic FEM. In the model, a new method was adopted to treat the friction behavior of the work-pieces as Coulomb friction law. and shear friction law at different inertia friction welding stages, respectively. The calculating method of friction torque, rotating velocity of the work-piece, input energy and friction heat was proposed on the base of above friction model. By coupling the subroutines of friction, friction torque, rotating velocity, input energy and friction heat into the FE model, the inertia friction welding process of GH4169 superalloy ring parts was successfully simulated. The distribution of temperature, stress and deformation of the joint and the evolutions of upset, rotating velocity, friction torque and input energy were calculated and analyzed. The corresponding experiments were carried out to validate the calculated results.A 3D rigid viscoplastic FE model of continuous drive friction welding was established with torsional friction in the friction interface being introduced into the calculation of stress and deformation. The model simulated two continuous drive friction-welding cases of GH4169 superalloy rod by performing coupled thermal and mechanical analyses. The temperature distribution, deformation and upset of the joint were calculated and validated by the experiments. Also, the deformation rate, the material flow and the possibility of defect occurrence in the joint were predicted and analyzed.The comparative study was carried out on the calculated results of temperature, deformation, stress, and upset by 3D model and 2D model, respectively. The results show that it is because of the introduction of trosional friction that the calculated equivalent stress of 3D model is larger, which lead to heavier deformation and bigger upset of the joint. Further comparison between the calculated and the experimental results show that the calculated results of 3D model fit the experimental results better and the 3D model can more accurately simulate the friction welding process. However there is only slight difference of calculated results between 3D model and 2D model, so the 3D model can be replaced by 2D model in the simulation of friction welding process.Applying the FE model of inertia friction welding, the influence of technological parameters and geometrical parameter of ring parts on the evolutions of upset and input energy was studied during inertia friction welding process. Several regulations were discovered: (1) increasing initial rotating velocity, increasing pressure or decreasing rotating inertia will produce an increase in upset, which is found to be most sensitive to initial rotating velocity, next to pressure, and most insensitive to rotating inertia. (2) Even keeping the input energy density constant in the inertia friction welding of different ring parts, the thickness of the ring parts has a significant influence on the upset, which will decrease with the ring thickness increasing. (3) At the initial stage of inertia friction welding, the technological parameters and geometrical parameters of ring parts have little influence on the input energy, while at the deformation period there is a significant influence both on quantity of input energy and energy input rate. With increasing of the pressure or decreasing of rotating inertia and ring thickness, the energy input rate will increase. At the same time if increasing the initial rotating velocity, only the quantity of input energy is enhanced, and the energy input rate nearly keep constant. All the above regulations have an important instruction in the optimization and investigation of friction welding parameters.Using the FE model of inertia friction welding, numerical simulations were carried out on the formation and extension of the plastic zone in the joint during entire inertia friction welding process. Several regulations were found: (1) the plastic zone covered in the joint as the shape of disk. (2) At the initial stage of inertia friction welding, the plastic zone is confined to a very thin layer. Then once the plastic zone begins to be produced, it will spread rapidly into the bulk of the work-piece material till the steady equilibrium stage of friction welding, when the plastic zone almost keep a constant thickness. (3) Only when the plastic zone extends to a critical thickness, the axial shortening of the joint can be produced. And wider the thickness of plastic zone, faster the upset rate happens.Based on the validated simulation results by the above model, an improved BP algorithm was used to construct the ANN prediction system of the upset according to technological parameters and geometrical parameter of the ring parts during inertia friction welding process. The upset was predicted using the trained ANN system, and perfect results were obtained.