【摘要】 为研究TC4钛合金的力学性能,借助Gleeble-1500D热模拟机,在等温、等速率条件下,对Ti-6Al-4V合金进行压缩。为进一步分析实验结果,将非线性建模方法高斯回归过程(GP)应用于TC4流变应力的预测,并提出相应的模型。实例研究表明,借助MATLAB语言编制程序,预测流变应力的高斯回归过程学习方法是科学、可行的,其绝对误差为0.91MPa,相对误差为4.84%,与ANN模型相比,预测精度更高,而且简单实用,为TC4热变形工艺的制定提供参考依据。更多还原

【Abstract】 In order to study the mechanical properties of Ti-6Al-4V alloy,the flow stress of Ti-6Al-4V was investigated under isothermal compression conditions with thermal simulation instrument(Gleeble-1500D).Gaussian process regression(GP) as nonlinear modeling approach was used for predicting the flow stress of TC4 and corresponding model was put forward.With the help of MATLAB,the GP model achieved reasonable prediction accuracy with mean Absolute error 0.91MPa and relative error 4.84%.Compared with ANN model,the GP model can forecast the flow stress more precisely and simply.It provides a reference for the formulation of the hot deformation process of TC4.更多还原