论文部分内容阅读
在Gleeble-1500热模拟实验机上对多组TC4钛合金试样进行热压缩实验,获得了变形温度在1053~1273 K、应变速率在0.01~10.00s-1情况下的真应力-应变曲线。通过BP神经网络对实验数据进行训练,建立了流变应力与应变、应变速率和温度相对应的预测模型,并对该模型的预测性能进行评估验证,采用预测数据构造了预测加工图,最后结合微观组织对预测加工图的可行性进行验证。结果表明,预测数据和实验数据的相关系数R为0.99886,平均相对误差为-0.21%,相对误差标准偏差为2.48 MPa,此模型具有良好的预测性能。预测加工图与实验加工图能够很好的吻合,通过预测加工图对材料的可加工性能进行预测,在一定程度上可以解决实验数据不足的缺陷。真应变为0.916的预测加工图大致分为A,B,C3个区域。失稳A区η值出现极小值(-0.16),应变速率较高时,材料局部发生动态再结晶,出现局部变形失稳的现象;应变速率较低时,组织很不均匀,易失稳。稳定B区具有较大的η值,并出现极大值(0.45),其α相球化效果显著、组织均匀,在相界处出现一定数量的细小等轴组织和较大比例的片状组织,确定此区为最优加工区。稳定C区α相球化效果比较明显,可作为加工区。
The hot compressive experiments of multiple samples of TC4 titanium alloy were carried out on a Gleeble-1500 thermal simulator. The true stress-strain curves of the samples were obtained at deformation temperatures between 1053 and 1273 K and strain rates between 0.01 and 10.00 s-1. The BP neural network is used to train the experimental data, and a predictive model of flow stress corresponding to strain, strain rate and temperature is established. The predictive performance of the model is evaluated and verified. Predictive processing diagrams are constructed by using predictive data. Finally, Micro-organizations to verify the feasibility of processing maps. The results show that the correlation coefficient between prediction and experimental data is 0.99886, the average relative error is -0.21%, and the relative standard deviation of error is 2.48 MPa. This model has good predictive performance. Predict the processing map and the experimental map can be a good match, by predicting the processing map of the material can be machined to predict the performance, to some extent, to solve the shortcomings of experimental data deficiencies. Really expected to become 0.916 forecast processing map is roughly divided into A, B, C3 area. When the strain rate is high, the material locally undergoes dynamic recrystallization and local deformation instability occurs. When the strain rate is low, the microstructure is very uneven and easily destabilized . Stable B region has a larger value of η, and the maximum value (0.45), the α-phase spheroidization effect was significant, uniform organization, in the phase boundary there is a certain number of small equiaxed and a larger proportion of flake , Determine this area is the best processing area. Stable C-phase α-phase spheroidization effect is obvious, can be used as processing area.