Font Size: a A A

Welding Process, Numerical Simulation Based On Finite Element Method And Artificial Neural Network

Posted on:2005-10-17Degree:MasterType:Thesis
Country:ChinaCandidate:G L LiangFull Text:PDF
GTID:2191360122497328Subject:Materials Processing Engineering
Abstract/Summary:PDF Full Text Request
Due to going through unequal heat and cooling process, metal components produced the welding residual stress and distortion during arc welding, which affected seriously resist brittle capacity, fatigue intensity, resist stress corrasion etc, furthermore, it also played down the rigidity and carrying capacity of welding components. Many factors have influences on the welding residual stress and distortion, such as welding current, welding voltage, welding speed, materials and thickness of workpieces. Therefore, optimizing welding parameters was a valid means to minish and control welding residual stress and distortion.In this paper, it was based on the physical tests and Finite Element Analysis (FEA) for AZ31magnesium alloy and 945steel plates, using Artificial Neural Network (ANN) to study the different parameters on welding residual stress and distortion during TIG butt-welding process. It also validated the welding temperature field, residual stress field and distortion field for AZ31 magnesium alloy and 945 steel plates by modern testing means. The study results presented that the welding residual stress would increase with the thickness or welding current increase and would decrease with welding speed increase, but the welding residual distortion would decrease with the thickness or welding speed increase and increase with the welding current increase. The welding residual stress and distortion using progressive welding would smaller than using continuous welding. Therefore, it can minish welding residual stress and distortion using little current, high speed and progressive welding in actual manufacture under ensuring complete penetration condition.In order to select the parameters quickly and efficiently, save test expenditure, improve welding speed and economic benefit, it developed a software system to predict the welding residual stress and distortion, which embedded the FEA and ANN. It not only can attain the biggest welding residual stress and distortion under different parameters, but also can attain dynamic curve map of the largest welding residual stress and distortion with different parameters, which had a certain directive sense.
Keywords/Search Tags:TIG butt-welding, Finite Element Analysis, Artificial Neural Network, Welding Residual Stress and Distortion, Welding Parameters Optimization
PDF Full Text Request
Related items