提出一种基于SABO-GRU-Attention(subtraction average based optimizer-gate recurrent unitattention)的锂电池SOC(state of charge)估计方法。采用基于平均减法优化算法自适应更新GRU神经网络的超参数,融合SE(squeeze and excitation...提出一种基于SABO-GRU-Attention(subtraction average based optimizer-gate recurrent unitattention)的锂电池SOC(state of charge)估计方法。采用基于平均减法优化算法自适应更新GRU神经网络的超参数,融合SE(squeeze and excitation)注意力机制自适应分配各通道权重,提高学习效率。对马里兰大学电池数据集进行预处理,输入电压、电流参数,进行锂电池充放电仿真实验,并搭建锂电池荷电状态实验平台进行储能锂电池充放电实验。结果表明,提出的SOC神经网络估计模型明显优于LSTM、GRU以及PSO-GRU等模型,具有较高的估计精度与应用价值。展开更多
针对现有的数字化档案多标签分类方法存在分类标签之间缺少关联性的问题,提出一种用于档案多标签分类的深层神经网络模型ALBERT-Seq2Seq-Attention.该模型通过ALBERT(A Little BERT)预训练语言模型内部多层双向的Transfomer结构获取进...针对现有的数字化档案多标签分类方法存在分类标签之间缺少关联性的问题,提出一种用于档案多标签分类的深层神经网络模型ALBERT-Seq2Seq-Attention.该模型通过ALBERT(A Little BERT)预训练语言模型内部多层双向的Transfomer结构获取进行文本特征向量的提取,并获得上下文语义信息;将预训练提取的文本特征作为Seq2Seq-Attention(Sequence to Sequence-Attention)模型的输入序列,构建标签字典以获取多标签间的关联关系.将分类模型在3种数据集上分别进行对比实验,结果表明:模型分类的效果F1值均超过90%.该模型不仅能提高档案文本的多标签分类效果,也能关注标签之间的相关关系.展开更多
This paper aims to investigate the role of bi-directional shear in the mechanical behaviour of granular materials and macro-micro relations by conducting experiments and discrete element method(DEM)modelling.The bi-di...This paper aims to investigate the role of bi-directional shear in the mechanical behaviour of granular materials and macro-micro relations by conducting experiments and discrete element method(DEM)modelling.The bi-directional shear consists of a static shear consolidation and subsequent shear under constant vertical stress and constant volume conditions.A side wall node loading method is used to exert bi-directional shear of various angles.The results show that bi-directional shear can significantly influence the mechanical behaviour of granular materials.However,the relationship between bidirectional shear and mechanical responses relies on loading conditions,i.e.constant vertical stress or constant volume conditions.The stress states induced by static shear consolidation are affected by loading angles,which are enlarged by subsequent shear,consistent with the relationship between bidirectional shear and principal stresses.It provides evidence for the dissipation of stresses accompanying static liquefaction of granular materials.The presence of bi-directional principal stress rotation(PSR)is demonstrated,which evidences why the bi-directional shear of loading angles with components in two directions results in faster dissipations of stresses with static liquefaction.Contant volume shearing leads to cross-anisotropic stress and fabric at micro-contacts,but constant vertical stress shearing leads to complete anisotropic stress and fabric at micro-contacts.It explains the differentiating relationship between stress-strain responses and fabric anisotropy under these two conditions.Micromechanical signatures such as the slip state of micro-contacts and coordination number are also examined,providing further insights into understanding granular behaviour under bi-directional shear.展开更多
文摘提出一种基于SABO-GRU-Attention(subtraction average based optimizer-gate recurrent unitattention)的锂电池SOC(state of charge)估计方法。采用基于平均减法优化算法自适应更新GRU神经网络的超参数,融合SE(squeeze and excitation)注意力机制自适应分配各通道权重,提高学习效率。对马里兰大学电池数据集进行预处理,输入电压、电流参数,进行锂电池充放电仿真实验,并搭建锂电池荷电状态实验平台进行储能锂电池充放电实验。结果表明,提出的SOC神经网络估计模型明显优于LSTM、GRU以及PSO-GRU等模型,具有较高的估计精度与应用价值。
文摘针对现有的数字化档案多标签分类方法存在分类标签之间缺少关联性的问题,提出一种用于档案多标签分类的深层神经网络模型ALBERT-Seq2Seq-Attention.该模型通过ALBERT(A Little BERT)预训练语言模型内部多层双向的Transfomer结构获取进行文本特征向量的提取,并获得上下文语义信息;将预训练提取的文本特征作为Seq2Seq-Attention(Sequence to Sequence-Attention)模型的输入序列,构建标签字典以获取多标签间的关联关系.将分类模型在3种数据集上分别进行对比实验,结果表明:模型分类的效果F1值均超过90%.该模型不仅能提高档案文本的多标签分类效果,也能关注标签之间的相关关系.
基金the funding support from National Natural Science Foundation of China(Grant No.42307243)Henan Province Science and Technology Research Project(Grant No.232102321102)Shanxi Provincial Key Research and Development Project(Grant No.202102090301009).
文摘This paper aims to investigate the role of bi-directional shear in the mechanical behaviour of granular materials and macro-micro relations by conducting experiments and discrete element method(DEM)modelling.The bi-directional shear consists of a static shear consolidation and subsequent shear under constant vertical stress and constant volume conditions.A side wall node loading method is used to exert bi-directional shear of various angles.The results show that bi-directional shear can significantly influence the mechanical behaviour of granular materials.However,the relationship between bidirectional shear and mechanical responses relies on loading conditions,i.e.constant vertical stress or constant volume conditions.The stress states induced by static shear consolidation are affected by loading angles,which are enlarged by subsequent shear,consistent with the relationship between bidirectional shear and principal stresses.It provides evidence for the dissipation of stresses accompanying static liquefaction of granular materials.The presence of bi-directional principal stress rotation(PSR)is demonstrated,which evidences why the bi-directional shear of loading angles with components in two directions results in faster dissipations of stresses with static liquefaction.Contant volume shearing leads to cross-anisotropic stress and fabric at micro-contacts,but constant vertical stress shearing leads to complete anisotropic stress and fabric at micro-contacts.It explains the differentiating relationship between stress-strain responses and fabric anisotropy under these two conditions.Micromechanical signatures such as the slip state of micro-contacts and coordination number are also examined,providing further insights into understanding granular behaviour under bi-directional shear.