基于人工神经网络的高精度Ti6242s合金热变形本构模型

使用Gleeble-3800对锻态Ti6242s钛合金在温度950～1010℃、应变速率0.01～10 s-1的条件下进行了75%变形量的热压缩模拟试验。基于实验取得的真应力-真应变曲线,分别使用人工神经网络(ANN)和Arrhenius方程建立Ti6242s合金本构模型,研究其热变形行为。结果表明:流变应力在变形开始后迅速上升至峰值应力,随后硬化与软化达到动态平衡,在真应变达到0.6后加工硬化逐渐占据主导,硬化幅度随应变速率的增大而提高;人工神经网络本构模型预测值的平均相对误差(AARE)为2.25%,决定系数(R2)为0.999 06;Arrhenius方程本构模型预测值的AARE为14.40%,R~2为0.954 68,精度在参数范围内波动较大;ANN本构模型精度远高于Arrhenius本构模型,且在整个参数范围内具有一致的精度;ANN本构模型具有良好的泛化能力,在实验参数范围外预测流变应力仍具有较高的精度。     

基于物理及元建模法的TB15钛合金本构模型研究

采用Gleeble-3800热模拟机对TB15钛合金进行等温恒应变速率热压缩试验,研究其在变形温度为810～930℃、应变速率为0.001～10s-1和高度压下量为60%条件下的热变形行为;建立了物理、支持向量回归(SVR)和响应面三种本构关系模型来预测TB15钛合金的流动应力,同时对比了三种本构模型的预测精度。结果表明:TB15钛合金的流动应力随应变速率的降低和变形温度的升高而减小,峰值应力的变化对应变速率的敏感性更高;物理本构模型、SVR本构模型和响应面本构模型相关系数R均大于0.98,但是响应面本构模型的R值达到了0.993,而且响应面本构模型的相对误差在±5%范围内的预测值频率达到了67.9%,大于物理本构模型的58.6%。同时经过方差分析得到所构建的响应面本构模型的显著性检验值P<0.0001,表明响应面本构模型预测的流动应力与变形温度、应变速率和应变之间的回归关系显著,比物理本构模型和SVR本构模型有更高的精度,能够更好的预测TB15钛合金的流动应力。     

基于人工神经网络的超细晶纯钛热变形本构模型研究

对等通道转角挤压(ECAP)制备的超细晶纯钛,在温度为250~450 ℃、应变速率为10_-5~1s_-1的条件下进行热压缩实验。基于真应力和真应变实验数据,分别使用人工神经网络(ANN)和Arrhenius方程建立超细晶纯钛的热变形本构模型,研究其热变形行为。实验结果表明：在变形初期,流变应力随应变的增大而升高,随后趋于平缓,最终流变应力达到一个稳定值。人工神经网络训练和预测结果表明：人工神经网络最佳结构为3×12×1,人工神经网络模型预测的平均相对误差(AARE)为2.1%,相关系数(R)为0.9979,Arrhenius方程模型预测的AARE为11.54%,R为0.9464。即人工神经网络模型能够更加精确的描述超细晶纯钛的本构关系。通过对比不同温度下两种模型的误差,人工神经网络模型在高温条件下具有更好的稳定性。     

基于神经网络的TC21合金本构关系模型(英文)

本构方程是描述材料变形和有限元模拟基本信息必要的数学模型,它反映流动应力与应变、应变率和温度综合作用的高度非线性关系。基于Gleeble-1500热模拟机上进行等温压缩试验获得的实验数据,系统研究TC21钛合金的流变行为,并采用BP人工神经网络建立该合金的本构关系模型。在该模型中,输入变量为应变、应变速率和变形温度,输出变量为流动应力。与传统方法相比,利用BP人工神经网络所建立的本构关系模型能够更好地表征试验数据及描述整个变形过程。     

Ti_2AlC/TiAl(Nb)复合材料热变形能力

采用Gleeble3500热模拟试验机对Ti2AlC/TiAl(Nb)复合材料进行高温压缩实验,实验温度范围为1000℃～1150℃,应变速率范围为10-3s-1～10-1s-1,工程压缩应变为50%,得到复合材料高温压缩真应力-真应变曲线。结果表明,Ti2AlC/TiAl(Nb)复合材料的高温变形流变应力对温度及应变速率敏感;流变应力随应变速率的增大而增大,随温度的升高而减小,可用位错-颗粒交互作用模型解释复合材料的应力-应变行为;Zenner-Hollomon参数的指数函数能够较好的描述该合金高温变形时的流变应力行为。建立的本构方程为ε=9.31×1011[sinh(0.0044σ)]2.52exp[-366.2/(RT)],其变形激活能为366.2kJ/mol。     

基于逐步回归法的TA15钛合金本构模型的建立

采用等温恒应变速率压缩实验研究,揭示了TA15钛合金热变形过程中变形温度、应变速率和应变对流动应力的影响规律.利用逐步回归法确定了影响流动应力的宏观可测变量(变形温度、应变速率和应变)的“最优”自变量子集,基于热压缩实验数据和部分神经网络预测数据,利用最小二乘估计算法求解模型回归参数值,建立了基于逐步回归法的TA15的材料本构模型.实验值与回归计算值的对比表明:所建立的材料本构模型描述TA15钛合金的热变形过程中所发生的动态软化过程(动态再结晶和动态回复)是合理可行的,且所建的材料本构模型形式简单,方便于应用到商用有限元软件所提供接口子程序中.     

(TiB_(2)+TiB)/Ti-22Al-25Nb复合材料的热变形行为与组织演变EI北大核心CSCD

以放电等离子烧结(TiB_(2)+TiB)增强Ti_(2)AlNb基复合材料为初始材料,在Gleeble-3800热模拟实验机上开展了(TiB_(2)+TiB)/Ti-22Al-25Nb复合材料的热压缩变形实验,研究了变形温度1060~1150℃、应变速率0.05~5 s^(-1)范围内复合材料的热变形行为。通过对流变应力-应变数据分析,构建了复合材料在B2单相区内的本构方程,分析了不同Zener-Hollomon(Z)参数下复合材料的组织演变规律。结果表明:(Ti B_(2)+TiB)/Ti-22Al-25Nb复合材料的峰值应力随变形温度的升高和应变速率的降低而降低,压缩曲线存在不连续屈服现象。Z值对复合材料的组织演变和变形机制均有重要影响。当ln Z值处于较高水平(35.88)时,复合材料出现局部塑性流动变形失稳区,动态再结晶程度较低,再结晶晶粒平均尺寸为3.82μm,增强颗粒粒径平均尺寸为6.93μm。当ln Z值处于较低水平(29.11~31.28)时,复合材料心部区域均发生完全动态再结晶。随着Z值降低,当ln Z为29.11时,动态再结晶晶粒长大,其平均尺寸增至9.16μm,并且由于B元素扩散的加快,促进了烧结残余TiB_(2)颗粒向Ti B晶须(Ti Bw)转变,原位反应更加充分,增强颗粒平均尺寸减小至2.77μm,TiBw的团簇现象明显减弱。     

TA15钛合金热变形应力应变曲线及本构模型

在变形温度分别为750,800,850,900,950,1000和1050℃,应变速率分别为0.001,0.01,0.1和1s~(-1)的条件下,对TA15钛合金进行了热压缩试验,分析了变形温度和应变速率对流动应力的影响。根据试验结果,计算了变形过程的温升,表明变形热所导致的温升大小与应变速率和应变均成正比,在T=750℃,ε=1s~(-1)的低温高应变速率条件下所产生的温升最大,可以达到122.63℃。基于Sellars-Tegart本构模型,建立了TA15钛合金热变形时的本构模型。     

基于BP神经网络的2D70铝合金本构关系模型

利用Thermecmastor-Z型热加工模拟试验机对2D70铝合金进行等温恒应变速率压缩试验,获得了不同变形温度、不同应变速率和不同真应变下的流动应力数据.结合实验数据和神经网络知识,建立了具有BP算法的人工神经网络,训练结束后的神经网络即成为2D70铝合金的一个知识基的本构关系模型.误差分析表明,该神经网络本构关系模型具有较高的精度,可用于指导2D70铝合金热加工工艺的制定,并可用于2D70铝合金热变形过程的有限元模拟.     

粉末冶金TiAl基合金高温变形行为及其本构模型

采用Gleeble-3800热模拟机研究粉末冶金Ti-47Al-2Cr-2Nb-0.2W-0.15B(摩尔分数,%)合金在变形温度为1 100～1 250 ℃、应变速率为10-3～100 s-1和变形率为50%条件下的高温变形行为.结果表明:Ti-47Al-2Cr-2Nb- 0.2W-0.15B合金在高温变形初始阶段,流动应力随应变的增加迅速增加;当应变超过一定值后,流变应力开始下降并逐渐趋于稳定,出现稳态流动特征;随着形变温度的升高和应变速率的增加,合金高温变形时的峰值应力和稳态应力显著降低.利用热模拟压缩实验数据,基于Arrhenius 方程和Zener-Hollomon参数,运用多元回归分析方法建立Ti-47Al-2Cr-2Nb-0.2W-0.15B合金在高温变形过程中的流变应力本构模型.应用DEFORMTM 3D软件验证该流变应力本构模型的有效性,结果表明所得高温流变应力本构模型能够较好地预测Ti-47Al-2Cr-2Nb-0.2W- 0.15B合金的高温变形行为.     

Flow behavior of Al–6.2Zn–0.70Mg–0.30Mn–0.17Zr alloy during hot compressive deformation based on Arrhenius and ANN models

The hot deformation behavior of Al–6.2Zn–0.70Mg–0.30Mn–0.17Zr alloy was investigated by isothermal compression test on a Gleeble–3500 machine in the deformation temperature range between 623 and 773 K and the strain rate range between 0.01 and 20 s^?1. The results show that the flow stress decreases with decreasing strain rate and increasing deformation temperature. Based on the experimental results, Arrhenius constitutive equations and artificial neural network (ANN) model were established to investigate the flow behavior of the alloy. The calculated results show that the influence of strain on material constants can be represented by a 6th-order polynomial function. The ANN model with 16 neurons in hidden layer possesses perfect performance prediction of the flow stress. The predictabilities of the two established models are different. The errors of results calculated by ANN model were more centralized and the mean absolute error corresponding to Arrhenius constitutive equations and ANN model are 3.49% and 1.03%, respectively. In predicting the flow stress of experimental aluminum alloy, the ANN model has a better predictability and greater efficiency than Arrhenius constitutive equations.     

物理基本构模型和BP人工神经网络模型预测AZ80镁合金高温流动应力的比较研究

基于变形温度250~400 ℃和应变速率0.001~1 s^-1条件下的铸态AZ80镁合金的热压缩试验数据,建立了基于应力位错关系和动态再结晶动力学的物理基本构模型以及前馈反向传播算法的人工神经网络（ANN）模型来预测AZ80镁合金的热变形行为。采用相关系数（R）、平均绝对相对误差（AARE）、相对误差（RE）3种统计学指标来验证2种模型的预测精度。结果表明,2种模型均可以准确预测AZ80镁合金的热变形行为。其中,ANN模型预测的应力值与实验数据更为吻合,其R和AARE分别为0.9991和2.02%,而物理基本构模型预测的R和AARE分别为0.9936和4.52%。ANN模型较好的预测能力归功于它擅长处理复杂的非线性关系,而物理基本构模型的预测能力是基于模型具有一定的物理意义,模型参数的确定充分考虑了热变形过程中的加工硬化（WH）、动态回复（DRV）和动态再结晶（DRX）的热动力学机制。最后,对这2种本构模型的优缺点及适用范围进行了比较讨论。     

Flow Stress Prediction of SiCp/Al Composites at Varying Strain Rates and Elevated Temperatures

With isothermal compression tests in the Gleeble-3500 system, the hot deformation behaviors of SiCp/Al composite were studied at a wide range of temperatures from 623 K to 773 K, and strain rates ranging from 0.001 s^?1 to 10 s^?1. Four different modeling methods such as the modified Zerilli-Armstrong model, the strain compensation Arrhenius-type model, the double multivariate nonlinear regression (DMNR) and the artificial neural model (ANN) were used to predict the flow stress. The suitability levels of these models were evaluated by contrasting both the correlation coefficient R _C and the average absolute relative error. The results show that the predictions of these four models can adequately meet the accuracy requirement according to the experimental data of this composite. With the increasing of the numbers of determined material constants and the complexity of computing methods, the predictability of these four methods is enhanced. The deformation parameters in the selected ranges such as strain rate and temperature have non-ignorable effect on predicted results of the previous two methods, while they have slight influence on DMNR and ANN.     

利用晶粒尺寸构建本构方程模拟AA1070铝的热变形行为

对原有Johnson-Cook本构模型中的项进行修正,提出一种新的现象学的、基于经验的本构模型.该模型可用于描述和预测具有不同初始晶粒尺寸的AA1070铝在热加工过程中的流变应力.该模型考虑热软化、应变速率硬化、应变硬化、初始晶粒尺寸及其相互影响,能够正确模拟具有不同应变、应变速率和初始晶粒尺寸AA1070铝的高温行为...     

Modeling Constitutive Relationship of Cu-0.4 Mg Alloy During Hot Deformation

For predicting the high-temperature deformation behavior in a Cu-0.4 Mg alloy, the true stress-strain data from isothermal hot compression tests on a Gleeble-1500 thermo-mechanical simulator, in a wide range of temperatures (500, 600, 700, 750, and 800 °C) and strain rates (0.005, 0.01, 0.1, 1, 5, and 10 s^?1), were employed to develop the Arrhenius-type constitutive model and the artificial neural network (ANN) constitutive model. Furthermore, prediction ability of the two models for high-temperature deformation behavior was evaluated. Correlation coefficients (R) between the experimental and predicted flow stress for the Arrhenius-type constitutive model and the ANN constitutive model are 0.9860 and 0.9998, respectively, and average absolute relative errors between the experimental and predicted flow stress for these two models are 5.3967% and 0.7401%, respectively. Results show that the ANN constitutive model can accurately predict the high-temperature deformation behavior over a wider range of temperatures and strain rates, while for the Arrhenius-type constitutive model there is greater divergence in the regime of high strain rates and low temperatures.     

Constitutive modeling of compression behavior of TC4 tube based on modified Arrhenius and artificial neural network models

Warm rotary draw bending provides a feasible method to form the large-diameter thin-walled(LDTW)TC4 bent tubes, which are widely used in the pneumatic system of aircrafts. An accurate prediction of flow behavior of TC4 tubes considering the couple effects of temperature,strain rate and strain is critical for understanding the deformation behavior of metals and optimizing the processing parameters in warm rotary draw bending of TC4 tubes. In this study, isothermal compression tests of TC4 tube alloy were performed from 573 to 873 K with an interval of 100 K and strain rates of 0.001, 0.010 and0.100 s~(-1). The prediction of flow behavior was done using two constitutive models, namely modified Arrhenius model and artificial neural network(ANN) model. The predictions of these constitutive models were compared using statistical measures like correlation coefficient(R), average absolute relative error(AARE) and its variation with the deformation parameters(temperature, strain rate and strain). Analysis of statistical measures reveals that the two models show high predicted accuracy in terms of R and AARE. Comparatively speaking, the ANN model presents higher predicted accuracy than the modified Arrhenius model. In addition, the predicted accuracy of ANN model presents high stability at the whole deformation parameter ranges, whereas the predictability of the modified Arrhenius model has some fluctuation at different deformation conditions. It presents higher predicted accuracy at temperatures of 573–773 K, strain rates of 0.010–0.100 s~(-1)and strain of 0.04–0.32, while low accuracy at temperature of 873 K, strain rates of 0.001 s~(-1)and strain of 0.36–0.48.Thus, the application of modified Arrhenius model is limited by its relatively low predicted accuracy at some deformation conditions, while the ANN model presents very high predicted accuracy at all deformation conditions,which can be used to study the compression behavior of TC4 tube at the temperature range of 573–873 K and the strain rate of 0.001–0.100 s~(-1). It can provide guideline for the design of processing parameters in warm rotary draw bending of LDTW TC4 tubes.     

轧制态ME20M镁合金热拉伸变形行为及加工图

在变形温度为623～773 K,应变速率为0.001～0.1 s~(-1)的条件下,通过INSPEKT Table 100 kN电子万能高温试验机对轧制态ME20M镁合金进行了热拉伸实验,分析了变形温度和应变速率对材料流动应力的影响,建立了热变形条件下的本构模型及加工图。结果表明:随着变形温度的降低和应变速率的升高,轧制态ME20M镁合金的流动应力增加;建立的本构模型预测峰值应力与实验结果吻合较好,平均相对误差为5.19%;考虑应变对本构模型中材料常数影响后的预测应力值与实验值的相关度较高,平均相对误差为6.00%;最佳热加工范围为673～773 K、应变速率0.001～0.01 s~(-1)。     

AA7075高强铝合金热冲压流变行为本构模型对比研究

在热冲压过程中,AA7075高强铝合金板料经充分固溶后移入室温模具进行冲压成形并淬火。为表征AA7075铝合金在热冲压工艺中的变形行为,在温度200~480℃、应变速率0.01~10s-1范围内进行了高温拉伸试验。基于Arrhenius类型本构模型、Johnson-Cook模型以及Zerilli-Armstrong模型提出了多种修正本构模型,并应用实验所获流变曲线进行了拟合。提出的修正模型通过将模型参数表示为应变、应变速率及温度相关的多项式函数耦合了应变、应变速率及温度对流变应力的影响,并通过均方误差(MSE)以及相关系数R值对模型流变应力预测准确性进行了评价。结果表明,修正的Johnson-Cook模型能够更加准确的预测AA7075高温流变行为。     

基于动态再结晶动力学方程的镍基高温合金热变形流变应力预测

采用等温热压缩实验,研究了一种典型镍基高温合金在1010-1160oC及0.001-1s-1条件下的高温流变行为。结果表明在合金的高温变形过程中发生了动态回复(DRV)以及动态再结晶(DRX)现象。通过深入分析不同变形条件下合金的高温流变行为,分别建立了合金在加工硬化-动态回复阶段以及动态再结晶阶段的流变应力本构方程。其中,在动态再结晶阶段,流变应力本构方程的建立是基于一种新型的动态再结晶动力学方程,该方程中引入了最大软化速率应变。此外,采用线性拟合的方法,建立了本构方程中材料常数与Zener-Hollomon参数间的函数关系。同时,通过对比分析流变应力的实测值和预测值,并计算两者之间的相关系数(R)和平均相对误差绝对值(AARE),验证了所建立本构方程的准确性,它可以精确预测所研究合金的高温流变应力。     

Role of activation energies of individual phases in two-phase range on constitutive equation of Zr–2.5Nb–0.5Cu alloy

Dominant phase during hot deformation in the two-phase region of Zr–2.5Nb–0.5Cu (ZNC) alloy was studied using activation energy calculation of individual phases. Thermo-mechanical compression tests were performed on a two-phase ZNC alloy in the temperature range of 700–925 °C and strain rate range of 10^?2–10 s^?1. Flow stress data of the single phase were extrapolated in the two-phase range to calculate flow stress data of individual phases. Activation energies of individual phases were then calculated using calculated flow stress data in the two-phase range. Comparison of activation energies revealed that α phase is the dominant phase (deformation controlling phase) in the two-phase range. Constitutive equations were also developed on the basis of the deformation temperature range (or according to phases present) using a sine-hyperbolic type constitutive equation. The statistical analysis revealed that the constitutive equation developed for a particular phase showed good agreement with the experimental results in terms of correlation coefficient (R) and average absolute relative error (AARE).