Traditional selection of combustion catalysis is time-consuming and labor-intensive.Theoretical calculation is expected to resolve this problem.The adsorption energy of HMX and O atoms on 13 metal oxides was calculate...Traditional selection of combustion catalysis is time-consuming and labor-intensive.Theoretical calculation is expected to resolve this problem.The adsorption energy of HMX and O atoms on 13 metal oxides was calculated using DMol3,since HMX and O are key substances in decomposition process.And the relationship between the adsorption energy of HMX,O on metal oxides(TiO_(2),Al_(2)O_(3),PbO,CuO,Fe_(2)O_(3),Co_(3)O_(4),Bi_(2)O_(3),NiO)and experimental T30 values(time required for the decomposition depth of HMX to reach 30%)was depicted as volcano plot.Thus,the T30 values of other metal oxides was predicted based on their adsorption energy on volcano plot and validated by previous experimental data.Further,the adsorption energy of HMX on ZrO_(2)and MnO_(2)was predicted based on the linear relationship between surface energy and adsorption energy,and T30 values were estimated based on volcano plot.The apparent activation energy data of HMX/MgO,HMX/SnO_(2),HMX/ZrO_(2),and HMX/MnO_(2)obtained from DSC experiments are basically consistent with our predicted T30 values,indicating that it is feasible to predict the catalytic activity based on the adsorption calculation,and it is expected that these simple structural properties can predict adsorption energy to reduce the large quantities of computation and experiment cost.展开更多
This article makes the first attempt in assessing the influence of active constrained layer damping(ACLD)treatment towards precise control of frequency responses of functionally graded skew-magneto-electroelastic(FGSM...This article makes the first attempt in assessing the influence of active constrained layer damping(ACLD)treatment towards precise control of frequency responses of functionally graded skew-magneto-electroelastic(FGSMEE)plates by employing finite element methods.The materials are functionally graded across the thickness of the plate in terms of modest power-law distributions.The principal equations of motion of FGSMEE are derived via Hamilton’s principle and solved using condensation technique.The effect of ACLD patches are modelled by following the complex modulus approach(CMA).Additionally,distinctive emphasis is laid to evaluate the influence of geometrical skewness on the attenuation capabilities of the plate.The accuracy of the current analysis is corroborated with comparison of previous researches of similar kind.Additionally,a complete parametric study is directed to understand the combined impacts of various factors like coupling fields,patch location,fiber orientation of piezoelectric patch in association with skew angle and power-law index.展开更多
Objective:To identify the cerebral activation patterns associated with the processes that occur during viewing Chinese and English words in native Chinese English learners. Methods:12 right-handed Chinese English lear...Objective:To identify the cerebral activation patterns associated with the processes that occur during viewing Chinese and English words in native Chinese English learners. Methods:12 right-handed Chinese English learners were divided into two groups equally,namely English majors and non-English majors,and took semantic judgement tasks of both English and Chinese words, for whom the fMRI images were collected.Results:To various degrees, all subjects demonstrated activation of associated cerebral regions in both hemispheres and the left hemisphere activation was more significant for most subjects. Except for classical regions involved in language processing,such as Wernicke areas and Broca areas,there were other activated cerebral regions, including cerebellum, limbic system and basal ganglia nucleus, etc. To sum up,there were apparent overlap for cerebral activation distribution and no specific processing areas for both tasks. The analysis of ROI(region of interest)suggested that subjects in specialized group were more dependent on right hemisphere to perform English words task. Conclusion:Language cognition is dominated by left hemisphere,which is also shared by the right hemisphere to various degrees and thus two hemispheres work by ways of both dissociation and coordination. It is possible that working strategy of the right hemisphere in English task is related to proficiency of the second language. A variety of distinctions are shared by each subject for language cognitive patterns.展开更多
This paper investigates the effect of porosity on active damping of geometrically nonlinear vibrations(GNLV)of the magneto-electro-elastic(MEE)functionally graded(FG)plates incorporated with active treatment constrict...This paper investigates the effect of porosity on active damping of geometrically nonlinear vibrations(GNLV)of the magneto-electro-elastic(MEE)functionally graded(FG)plates incorporated with active treatment constricted layer damping(ATCLD)patches.The perpendicularly/slanted reinforced 1-3 piezoelectric composite(1-3 PZC)constricting layer.The constricted viscoelastic layer of the ATCLD is modeled in the time-domain using Golla-Hughes-Mc Tavish(GHM)technique.Different types of porosity distribution in the porous magneto-electro-elastic functionally graded PMEE-FG plate graded in the thickness direction.Considering the coupling effects among elasticity,electrical,and magnetic fields,a three-dimensional finite element(FE)model for the smart PMEE-FG plate is obtained by incorporating the theory of layer-wise shear deformation.The geometric nonlinearity adopts the von K arm an principle.The study presents the effects of a variant of a power-law index,porosity index,the material gradation,three types of porosity distribution,boundary conditions,and the piezoelectric fiber’s orientation angle on the control of GNLV of the PMEE-FG plates.The results reveal that the FG substrate layers’porosity significantly impacts the nonlinear behavior and damping performance of the PMEE-FG plates.展开更多
构造一种适用于反向传播(backpropagation,BP)神经网络的新型激活函数Lfun(logarithmic series function),并使用基于该函数的BP神经网络进行机床能耗状态的预测。首先,分析Sigmoid系列和ReLU系列激活函数的特点和缺陷,结合对数函数,构...构造一种适用于反向传播(backpropagation,BP)神经网络的新型激活函数Lfun(logarithmic series function),并使用基于该函数的BP神经网络进行机床能耗状态的预测。首先,分析Sigmoid系列和ReLU系列激活函数的特点和缺陷,结合对数函数,构造了一种非线性分段含参数激活函数。该函数可导且光滑、导数形式简单、单调递增、输出均值为零,且通过可变参数使函数形式更灵活;其次,通过数值仿真实验在公共数据集上将Lfun函数与Sigmoid、ReLU、tanh、Leaky_ReLU和ELU函数的性能进行对比;最后,使用基于Lfun函数的BP神经网络进行机床能耗状态的预测。实验结果表明,使用Lfun函数的BP神经网络相较于使用其他几种常用激活函数的网络具有更好的性能。展开更多
现有目标检测算法对背景复杂下小交通标志的检测效果并不理想。为此,提出了一种基于归一化通道注意力机制YOLOv7的交通标志检测算法(YOLOv7 based on normalized channel attention mechanism,YOLOv7-NCAM)。为了使YOLOv7-NCAM模型具有...现有目标检测算法对背景复杂下小交通标志的检测效果并不理想。为此,提出了一种基于归一化通道注意力机制YOLOv7的交通标志检测算法(YOLOv7 based on normalized channel attention mechanism,YOLOv7-NCAM)。为了使YOLOv7-NCAM模型具有像素级建模能力,提高它对小目标交通标志特征的提取能力,YOLOv7-NCAM算法使用FReLU激活函数构建了DBF和CBF两种卷积层,并用它们来组建模型的Backbone模块和Neck模块;提出一种归一化通道注意力机制(normalized channel attention mechanism,NCAM)并加入Head模块中。通过与整体网络一起训练,得到归一化(batch normalization,BN)缩放因子,利用缩放因子算出各个通道的权重因子,提升网络对交通标志特征的表达能力,从而使YOLOv7-NCAM网络模型能够集中关注检测目标交通标志。通过在CCTSDB-2021交通标志检测数据集上的测试,与YOLOv7网络模型对比结果表明,YOLOv7-NCAM算法对背景复杂下小交通标志的检测各项指标均有明显提高:准确率(precision,P)达到91.5%,比原网络高出9.5个百分点;召回率(recall,R)达到85.9%,比原网络高出5.7个百分点;均值平均精度(mean average precision,mAP)达到了91.4%,比原网络高出4.7个百分点。与现有的交通标志检测算法相比,YOLOv7-NCAM算法的检测准确率也有提高,且检测速度48.3 FPS,能满足实时需求。展开更多
基金supported by Key Science and Technology Innovation Team of Shaanxi Province(No.2022TD-33)National Natural Science Foundation of China(Grant Nos.21373161,21504067)。
文摘Traditional selection of combustion catalysis is time-consuming and labor-intensive.Theoretical calculation is expected to resolve this problem.The adsorption energy of HMX and O atoms on 13 metal oxides was calculated using DMol3,since HMX and O are key substances in decomposition process.And the relationship between the adsorption energy of HMX,O on metal oxides(TiO_(2),Al_(2)O_(3),PbO,CuO,Fe_(2)O_(3),Co_(3)O_(4),Bi_(2)O_(3),NiO)and experimental T30 values(time required for the decomposition depth of HMX to reach 30%)was depicted as volcano plot.Thus,the T30 values of other metal oxides was predicted based on their adsorption energy on volcano plot and validated by previous experimental data.Further,the adsorption energy of HMX on ZrO_(2)and MnO_(2)was predicted based on the linear relationship between surface energy and adsorption energy,and T30 values were estimated based on volcano plot.The apparent activation energy data of HMX/MgO,HMX/SnO_(2),HMX/ZrO_(2),and HMX/MnO_(2)obtained from DSC experiments are basically consistent with our predicted T30 values,indicating that it is feasible to predict the catalytic activity based on the adsorption calculation,and it is expected that these simple structural properties can predict adsorption energy to reduce the large quantities of computation and experiment cost.
文摘This article makes the first attempt in assessing the influence of active constrained layer damping(ACLD)treatment towards precise control of frequency responses of functionally graded skew-magneto-electroelastic(FGSMEE)plates by employing finite element methods.The materials are functionally graded across the thickness of the plate in terms of modest power-law distributions.The principal equations of motion of FGSMEE are derived via Hamilton’s principle and solved using condensation technique.The effect of ACLD patches are modelled by following the complex modulus approach(CMA).Additionally,distinctive emphasis is laid to evaluate the influence of geometrical skewness on the attenuation capabilities of the plate.The accuracy of the current analysis is corroborated with comparison of previous researches of similar kind.Additionally,a complete parametric study is directed to understand the combined impacts of various factors like coupling fields,patch location,fiber orientation of piezoelectric patch in association with skew angle and power-law index.
文摘Objective:To identify the cerebral activation patterns associated with the processes that occur during viewing Chinese and English words in native Chinese English learners. Methods:12 right-handed Chinese English learners were divided into two groups equally,namely English majors and non-English majors,and took semantic judgement tasks of both English and Chinese words, for whom the fMRI images were collected.Results:To various degrees, all subjects demonstrated activation of associated cerebral regions in both hemispheres and the left hemisphere activation was more significant for most subjects. Except for classical regions involved in language processing,such as Wernicke areas and Broca areas,there were other activated cerebral regions, including cerebellum, limbic system and basal ganglia nucleus, etc. To sum up,there were apparent overlap for cerebral activation distribution and no specific processing areas for both tasks. The analysis of ROI(region of interest)suggested that subjects in specialized group were more dependent on right hemisphere to perform English words task. Conclusion:Language cognition is dominated by left hemisphere,which is also shared by the right hemisphere to various degrees and thus two hemispheres work by ways of both dissociation and coordination. It is possible that working strategy of the right hemisphere in English task is related to proficiency of the second language. A variety of distinctions are shared by each subject for language cognitive patterns.
文摘This paper investigates the effect of porosity on active damping of geometrically nonlinear vibrations(GNLV)of the magneto-electro-elastic(MEE)functionally graded(FG)plates incorporated with active treatment constricted layer damping(ATCLD)patches.The perpendicularly/slanted reinforced 1-3 piezoelectric composite(1-3 PZC)constricting layer.The constricted viscoelastic layer of the ATCLD is modeled in the time-domain using Golla-Hughes-Mc Tavish(GHM)technique.Different types of porosity distribution in the porous magneto-electro-elastic functionally graded PMEE-FG plate graded in the thickness direction.Considering the coupling effects among elasticity,electrical,and magnetic fields,a three-dimensional finite element(FE)model for the smart PMEE-FG plate is obtained by incorporating the theory of layer-wise shear deformation.The geometric nonlinearity adopts the von K arm an principle.The study presents the effects of a variant of a power-law index,porosity index,the material gradation,three types of porosity distribution,boundary conditions,and the piezoelectric fiber’s orientation angle on the control of GNLV of the PMEE-FG plates.The results reveal that the FG substrate layers’porosity significantly impacts the nonlinear behavior and damping performance of the PMEE-FG plates.
文摘构造一种适用于反向传播(backpropagation,BP)神经网络的新型激活函数Lfun(logarithmic series function),并使用基于该函数的BP神经网络进行机床能耗状态的预测。首先,分析Sigmoid系列和ReLU系列激活函数的特点和缺陷,结合对数函数,构造了一种非线性分段含参数激活函数。该函数可导且光滑、导数形式简单、单调递增、输出均值为零,且通过可变参数使函数形式更灵活;其次,通过数值仿真实验在公共数据集上将Lfun函数与Sigmoid、ReLU、tanh、Leaky_ReLU和ELU函数的性能进行对比;最后,使用基于Lfun函数的BP神经网络进行机床能耗状态的预测。实验结果表明,使用Lfun函数的BP神经网络相较于使用其他几种常用激活函数的网络具有更好的性能。
文摘现有目标检测算法对背景复杂下小交通标志的检测效果并不理想。为此,提出了一种基于归一化通道注意力机制YOLOv7的交通标志检测算法(YOLOv7 based on normalized channel attention mechanism,YOLOv7-NCAM)。为了使YOLOv7-NCAM模型具有像素级建模能力,提高它对小目标交通标志特征的提取能力,YOLOv7-NCAM算法使用FReLU激活函数构建了DBF和CBF两种卷积层,并用它们来组建模型的Backbone模块和Neck模块;提出一种归一化通道注意力机制(normalized channel attention mechanism,NCAM)并加入Head模块中。通过与整体网络一起训练,得到归一化(batch normalization,BN)缩放因子,利用缩放因子算出各个通道的权重因子,提升网络对交通标志特征的表达能力,从而使YOLOv7-NCAM网络模型能够集中关注检测目标交通标志。通过在CCTSDB-2021交通标志检测数据集上的测试,与YOLOv7网络模型对比结果表明,YOLOv7-NCAM算法对背景复杂下小交通标志的检测各项指标均有明显提高:准确率(precision,P)达到91.5%,比原网络高出9.5个百分点;召回率(recall,R)达到85.9%,比原网络高出5.7个百分点;均值平均精度(mean average precision,mAP)达到了91.4%,比原网络高出4.7个百分点。与现有的交通标志检测算法相比,YOLOv7-NCAM算法的检测准确率也有提高,且检测速度48.3 FPS,能满足实时需求。
