TB8钛合金固溶组织研究及神经网络预测

Research and ANN Prediction on the Microstructure after Solution Treatment of TB8 Titanium Alloy

摘要:

深入分析了各变形工艺参数对TB8合金固溶处理显微组织的影响规律，建立了固溶组织再结晶体积分数、平均晶粒尺寸与变形工艺参数间的神经网络预测模型。结果表明，冷却和热处理制度相同的条件下，变形温度、变形程度和应变速率等变形工艺参数对TB8钛合金形变且固溶处理后的显微组织有重要的影响，若想获得晶粒较为细小且均匀的组织，需要在合适的应变速率下适当提高变形程度和降低变形温度；人工神经网络的预测结果与实测结果的高度拟合，表明人工神经网络模型可以较为精确地预测TB8合金的显微组织随变形工艺参数的变化而变化的情况。以上研究工作为TB8合金热加工工艺的制定提供了更为科学的理论依据

Abstract:

Analysis on the influence of deformation temperature, strain rate and deformation degree on the microstructure after solution treatment was carried out, and a predicting model for the recrystallization volume and average grain size of the microstructure was established by a three-layer feed-forward artificial neural network with a back-propagation learning rule. The results indicate that under the condition of the same cooling and heat treatment rules, the deformation parameters such as deformation temperature, deformation degree and strain rate have important influence on the microstructures evolution after hot deformation and solution treatment of TB8 titanium alloy. A larger deformation degree, a lower temperature and an appropriate strain rate are required to acquire the microstructure with homogeneous and fine grains. The close agreement of the predicted results with measured ones shows that the neural network is able to successfully predict the variation of the microstructure with the hot deformation parameters. The above results can provide the determining of reasonable hot forming process with a scientific base.