The void evolution equation and the elastoplastic constitutive model of casting magnesium alloy were investigated. The void evolution equation consists of the void growth and the void nucleation equations. The void gr...The void evolution equation and the elastoplastic constitutive model of casting magnesium alloy were investigated. The void evolution equation consists of the void growth and the void nucleation equations. The void growth equation was obtained based on the continuous supposition of the material matrix,and the void nucleation equation was derived by assuming that the void nucleation follows a normal distribution. A softening function related to the void evolution was given. After the softening function was embedded to a nonclassical elastoplastic constitutive equation,a constitutive model involving void evolution was obtained. The numerical algorithm and the finite element procedure related to the constitutive model were developed and applied to the analysis of the distributions of the stress and the porosity of the notched cylindrical specimens of casting magnesium alloy ZL305. The computed results show satisfactory agreement with the experimental data.展开更多
假定模型参数的不确定性服从正态分布,根据贝叶斯原理,其最可能的分布是结合先验信息和观测信息得到的最大后验概率,马尔科夫链蒙特卡罗(Markov Chain Monte Carlo,MCMC)抽样适用于此类反问题求解。鉴于随机论方法的巨大计算量,本研究利...假定模型参数的不确定性服从正态分布,根据贝叶斯原理,其最可能的分布是结合先验信息和观测信息得到的最大后验概率,马尔科夫链蒙特卡罗(Markov Chain Monte Carlo,MCMC)抽样适用于此类反问题求解。鉴于随机论方法的巨大计算量,本研究利用BP(Back Propagation)神经网络及相对熵最小化来自适应加密训练数据,从而建立替代复杂正向程序的代理模型,并利用开发的不确定性分析程序对影响空泡份额的模型参数不确定性进行量化分析,选用的子通道程序为COBRA-IV。结果表明:在求得模型参数不确定性后,通过不确定性正向传递得到结果的95%置信区间对实验值的包络性较好,利用不确定性均值对模型进行标定得到的结果较基准值更接近实验值。因此,本研究建立的不确定性量化分析方法能较好适用于子通道程序的不确定性分析。展开更多
基金Project(10572157) supported by the National Natural Science Foundation of China
文摘The void evolution equation and the elastoplastic constitutive model of casting magnesium alloy were investigated. The void evolution equation consists of the void growth and the void nucleation equations. The void growth equation was obtained based on the continuous supposition of the material matrix,and the void nucleation equation was derived by assuming that the void nucleation follows a normal distribution. A softening function related to the void evolution was given. After the softening function was embedded to a nonclassical elastoplastic constitutive equation,a constitutive model involving void evolution was obtained. The numerical algorithm and the finite element procedure related to the constitutive model were developed and applied to the analysis of the distributions of the stress and the porosity of the notched cylindrical specimens of casting magnesium alloy ZL305. The computed results show satisfactory agreement with the experimental data.
文摘假定模型参数的不确定性服从正态分布,根据贝叶斯原理,其最可能的分布是结合先验信息和观测信息得到的最大后验概率,马尔科夫链蒙特卡罗(Markov Chain Monte Carlo,MCMC)抽样适用于此类反问题求解。鉴于随机论方法的巨大计算量,本研究利用BP(Back Propagation)神经网络及相对熵最小化来自适应加密训练数据,从而建立替代复杂正向程序的代理模型,并利用开发的不确定性分析程序对影响空泡份额的模型参数不确定性进行量化分析,选用的子通道程序为COBRA-IV。结果表明:在求得模型参数不确定性后,通过不确定性正向传递得到结果的95%置信区间对实验值的包络性较好,利用不确定性均值对模型进行标定得到的结果较基准值更接近实验值。因此,本研究建立的不确定性量化分析方法能较好适用于子通道程序的不确定性分析。