A strategy on the consistency of tensile strength of friction stir lap welding joint based on the same peak temperature

The strength at the end of weld is relatively small due to the sharp increase in welding temperature, resulting in a problem of inconsistent joint strength along the weld direction. Although this problem can be solved by means of placing the lead-out plate or removing the weld end, it inevitably leads to the waste of materials. In view of this, this study proposed a technical strategy for obtaining the equal peak temperature based on the synergy of numerical simulation and artificial neural network model, and thereby fabricating the welded joint with consistent strength along the weld direction. The friction stir lap welding (FSLW) process of 2024-T4 aluminum alloys was taken as the research object. Firstly, the samples for radial basis function neural network (RBFNN) were obtained by the numerical simulation method. Then, taking the rotating speed, the welding speed at the steady stage and the welding speed at the ending stage as the input, and the difference in the peak temperatures between these two stages as the output, the RBFNN prediction model was established, and then the regulation scheme of the welding speed at the ending stage was obtained based on the idea of equal peak temperature. Finally, the reasons for the consistency of joint strength along the weld direction were further investigated by the changes in the hook morphology at the advancing side and the microhardness of the material. The research results showed that based on the samples obtained by numerical simulation, the prediction accuracy of RBFNN model exceeded 99%. By comparing the joint strengths at the ending and steady stages, the regulation of the welding speed at the ending stage reduced the strength fluctuation of FSLW joint from 28.8% to 4.3%. • The strength consistency control strategy of welded joint was proposed based on equal peak temperature. • Intelligent regulation of process parameters was realized through the integration of numerical simulation and artificial neural networks. • The influence mechanism of temperature consistency on joint formation and tensile strength was revealed.