基于BP神经网络的TC11钛合金工艺-性能模型预测

材料工艺与性能的关系具有复杂、非线性交互等特点。本文根据TC11钛合金力学性能与其影响因素之间的映射关系,以大量的试验数据为基础,建立了BP神经网络模型。模型的输入包括锻造温度、锻后冷却方式等热加工工艺参数;输出为常用的力学性能指标,即抗拉强度、屈服强度、延伸率和断面收缩率。运用该模型对TC11钛合金力学性能进行了预测,并通过试验数据对模型的预测精度进行了可靠性验证。同时,运用已建立的神经网络模型对TC11钛合金工艺参数与力学性能的关系进行了分析。结果表明,所建立的力学性能预测模型具有良好的外推能力,并且可以很好地反映出该合金的工艺-性能之间的复杂关系。

Model Prediction of Processing-Property of TC11 Titanium Alloy Using Artificial Neural Network

The relationship between processing and property of materials is complex. In the present investigation, based on a lot of experimental data, the technique of artificial neural network was employed to develop the prediction model of processing and property for TC11 titanium alloy. The inputs of the neural network were different forging process parameters such as forging temperature, forging style and cooling style. The outputs of the model were the tensile properties, including ultimate tensile strength, yield strength, elongation and reduction of area. The mechanical properties of TC11 titanium alloy were predicted by the established model, and the accuracy of the prediction was compared with the experimental data. Besides, the model was used to study the influence of the processing on the properties of TC11 titanium alloy. Results show that the model can predict the properties of this alloy with high accuracy and reliability, and the complex relationship between processing and properties can be well presented by the trained neural network, which is consistent with the metallurgical trends