摘要
6061铝合金在热变形时稳态应力主要受应变、温度、应变速率的影响。为了提高稳态应力的预测精度,基于Sellars-Tegart本构方程和BP神经网络建立了6061铝合金的预测模型,对比和分析了两个模型的预测效果。结果表明,两个预测模型得到的预测值均与试验值吻合程度较高,可较好地描述稳态应力与各热力参数之间的非线性关系,而且通过决定系数和标准残差的对比证实了BP神经网络预测模型具有更高的预测精度。
The steady state stress of 6061 aluminum alloy during deformation is mainly affected by strain, temperatureand strain rate. In order to improve the prediction accuracy of steady state stress, the prediction models of 6061 aluminumalloy were established based on Sellars-Tegart constitutive equation and BP neural network, and the predicted effects of thetwo models were compared and analyzed. The results indicate that the predicted values obtained by the two prediction modelsare in agreement with the experimental values, and the nonlinear relationship between steady state stress and thermalparameters can be well described. Furthermore, it is confirmed by comparing determination coefficients and standard residualsthat the prediction model of BP neural network has higher prediction accuracy.
出处
《热加工工艺》
CSCD
北大核心
2016年第21期131-134,共4页
Hot Working Technology
基金
国家自然科学基金资助项目(11572187)
上海市教育委员会科研创新项目(14ZZ154)
上海市科学技术委员会项目(14DZ2261000
13160501000)
关键词
6061
铝合金
稳态应力
本构方程
BP
神经网络
预测精度
6061 aluminum alloy
steady state stress
constitutive equation
BP neural network
predictive precision
