论文部分内容阅读
在预设实验条件下,利用Gleeble-3500D热模拟机,完成了钛合金TC4高温超塑性拉伸试验。然后利用处理高度非线性问题的高斯回归技术,借助MATLAB语言编程,对高温超塑性拉伸过程中的流变应力进行了预测,其平均绝对误差0.67 MPa,平均相对误差2.91%。与神经网络预测结果相比,其预测精度更高且简单易行,是钛合金超塑性变形过程中参数预测和优化的可行工具。
Under the preset experimental conditions, the high temperature superplastic tensile test of titanium alloy TC4 was completed by Gleeble-3500D thermal simulator. Then using the Gaussian regression technique to deal with highly nonlinear problems, the flow stress in the process of high temperature superplastic stretching was predicted by MATLAB language programming. The average absolute error was 0.67 MPa and the average relative error was 2.91%. Compared with the prediction results of neural network, the prediction accuracy is higher and simple and feasible, which is a feasible tool to predict and optimize the parameters in superplastic deformation of titanium alloy.