| Friction Stir Welding (FSW) is a solid state welding method which developed recently. From the beginning of the 21st century, depending on the peculiar advantage, FSW is paid much attention in the welding field. Especial for the aluminums alloys materials which difficultly welded by the normal welding methods, it would not appear the defects such as cracks, voids and shrinkages when using the FSW method. At the present time, most of the researches are focused on the FSW of aluminum alloys. The FSW theory has become more mature, and the welding was gradually promoted to the aviation industry.The welding-joint's fatigue behavior of FSW is one of the evaluations for the invalidation behavior to the FSW structure-pieces. However, at the present time there are few researches for the fatigue capability on the FSW of aluminum alloys, this is the key factor which hinders the development of FSW. As a result, the fatigue capability researches on the FSW of aluminum alloys are actually and engineering significations.In this paper, the suitable FSW welding-parameters for four aluminum alloys (7075 7050 6082 2024) were ascertained in the welding experiment. The relationships between FSW welding-parameters and weld fatigue life were analyzed linked to the result of the fatigue test. By the data which acquired in the experiment, the prediction model for the aluminum alloys FSW high and low fatigue life was established based on the a artificial neural network (ANN). The back propagation (BP) neural network models, radial basis function (RBF) neural network models and the Graphical User Interface were established by the using of MATLAB language. When using the training-over network to predict the FSW joints fatigue life, Comparing with the experiment data, it shows that the prediction data were tally well with the experiment data, fulfilled with the demand of the engineering. The network provides a new way to evaluate and forecast the strength of the FSW joints fatigue life. |
PDF Full Text Request