基于模糊神经网络的Zr-2合金晶粒尺寸及流变应力模型

A Model for Grain Size and Flow Stress of Zi-2 Alloy Based on Fuzzy Neural Network

通过Zr-2合金在变形温度为750,800℃,变形速率为0.01,1,10s-1,变形程度为50%,65%的热压缩试验获得数据,采用模糊神经网络方法建立Zr-2合金的晶粒尺寸及流变应力模型。模型的输入参数包括变形温度、变形程度、变形速率等热加工参数,模型的输出为晶粒尺寸和流变应力。结果表明,该模型避免了传统经验回归拟合复杂的数学公式计算,是简单而精确的建模方法,可用于优化热加工参数。

A modular fuzzy neural network model is proposed to predict the flow stress and grain size of Zr-2 alloy in hot deformation. The input parameters of the fuzzy neural network model are deformation temperature, deformation degree and strain rate; the outputs of the model are flow stress and grain size. Compared with the traditional regression method, the fuzzy neural network model have the simple and accurate characteristics, and can be used to optimize hot processing parameters.