基于人工神经网络的钛合金焊接接头机械性能预测

Mechanical Property Prediction of Welding Joint Based on Artificial Neural Network for Titanium Alloy

采用多层前向反馈神经网络模型,对钛合金钨极氩弧焊的焊接接头机械性能进行了模拟和预测。其中,输入参数包括钛合金成分、冷却速度和热处理参数;输出参数包括5个重要的机械性能,即极限抗拉强度、延伸率、断面收缩率、屈服强度和硬度。详细分析了铝和钒这2种元素对机械性能的影响。

The mechanical properties of gas tungsten arc welding （GTAW） are simulated predicted by multilayer forward neural network model for titanium alloys. The input parameters of the neural network are alloy compositions, cooling rate and heat treatment conditions, and the output parameters of the neural network are five important mechanical properties of the weld metal of titanium alloys, namely ultimate tensile strength, elongation, reduction of area, yield strength and hardness. The effects of aluminum and vanadium on mechanical properties were investigated in detail.