الفهرس | Only 14 pages are availabe for public view |
Abstract In this work bobbin tool friction stir welding (BT-FSW) process was used to produce lap joint of total thickness of 10 mm out of 5mm thick AA 1050 sheets. Bobbin tool with different pin geometries (cylindrical, square, and triangular) and concave shoulder profile were manufactured and used in BT-FSW of lap joints at various travel speeds, and at 600 rpm rotation rate. The weld joints were evaluated using visual inspection, and macrostructure investigation, as well as tensile and hardness testing. A defect free welds were produced using bobbin tool with cylindrical, and square pin profile at all parameters. Defect-free BT-FSW joints with relatively high tensile strength, using a square pin tool were produced. BT-FSW at a travel speed 200 mm/min leads to better tensile properties, in case of using square pin. In terms of hardness results, the increase of travel speed results in an increase of the hardness in the middle zone of cross-section area. A mathematical model for heat generation in BT-FSW of AA1050 using different pin profiles were developed. Using a mathematical approach, it is seen that by increasing the number of edges from 3 for triangular to 4 edges for square pin, the amount of heat generation initially increases from the Triangular (Tr) to square (Sq) pin profile, then decreases to the cylindrical pin profile. With the proposed mathematical approach, one can directly predict the weld temperature for respective tool geometry under given process parameters, which will be helpful for predicting the mechanical properties for that aluminum alloy. Artificial Neural Networks (ANN) modeling is developed to predict the effect of bobbin tool friction stir welding (BT-FSW) on tensile shear load of AA1050. Results indicate that the networks prediction is very closed to the experiment results. Overall correlation coefficient (R) value for training, validation and testing of ANN model is higher than 0.99. |