基于BP神经网络的TC4钛合金超塑性变形后组织及性能预测研究

Predicting Mechanical Properties and Microstructure of TC4 Alloy after Superplastic Forming based on BP Artificial Neural Network

分别用VisualFortran语言和MATLAB软件建立了TC4钛合金超塑性变形时变形参数与其力学性能和晶粒尺寸之间的BP神经网络模型,通过用较少的力学性能和晶粒尺寸的试验数据进行训练,进而对其性能进行预测。结果表明,BP神经网络用于材料超塑性变形后的力学性能及晶粒尺寸预测是可行的,其预测误差小于7%。

Based on BP artificial neural network, an investigation was carried out to predict mechanical properties and microstructure of TC4 alloy after superplastic forming. A BP artificial neural network model between superplastic deformation parameters and mechanical properties and microstructure of TC4 has been built, using Visual Fortran language and on the basis of Neural Network Toolbox in MATLAB software. It is trained for predicting the performance of TC4 alloy after superplastic deformation. The results show that BP artificial neural network can be used in predicting mechanical properties and grain size of materials after superplastic forming and its predicting error is less than 7%.