基于模糊神经网络的TC4-DT钛合金高温变形本构关系模型(英文)

Modeling the High Temperature Deformation Constitutive Relationship of TC4-DT Alloy Based on Fuzzy-neural Network

通过分析研究变形温度、应变速率及变形程度参数对TC4-DT钛合金高温变形行为的影响，建立了一种基于自适应模糊神经网络的TC4-DT钛合金高温变形本构关系预测模型。高温变形热模拟压缩试验的变形温度为750～1150℃，应变率为O．001～10s^（-1），试样高度压缩率为50％。本研究中建立的网络模型集成了模糊推理系统误差反向传播（BP）神经网络的学习算法。结果表明，该模型的预测值与实验结果比较吻合，最大相对误差小于6％。本研究证明模糊神经网络是一种优化TC4-DT钛合金本构关系模型和优化变形工艺参数的有效、实用方法。

By analyzing the high temperature TC4-DT titanium alloys＇ deformation temperature, strain rate and deformation degree with the parameters of the experimental data flow stress, an adaptive fuzzy-neural network model has been established to predict flow stress data to model the high temperature deformation constitutive relationship of TC4-DT titanium alloy. The experimental results were obtained at deformation temperature of 750-1150 ℃, strain rates of 0.001- 10 s1, and height reduction of 50%. The network integrates the fuzzy inference system with a back-propagation （BP） learning algorithm of neural network. Results show that the predicated values are in satisfactory agreement with the experimental results and the maximum relative error is less than 6%. It proves that the fuzzy-neural network is a very effective and practical method to achieve more optimized TC4 - DT titanium alloy constitutive relation model and optimize deformation process parameters.